Psychoactive substances vary considerably in their abuse liability. For example, major tranquilizers typically do not maintain self-administration behavior, whereas smokable forms of nicotine and cocaine are highly addictive. Quantification of the differences among drugs in terms of abuse liability and discovering the mechanisms that underlie these differences is fundamental in the development of safer medications in general and in developing more effective medications for the treatment of addictive disorders. Two strategies for human evaluation of the mechanisms of drug abuse liability are drug discrimination and drug self-administration. Both of these paradigms evolved from animal research, and their application to humans makes it possible to apply animal and human data to identify mechanisms of addiction. Currently, we are using the drug discrimination paradigm to explore the mechanisms underlying the distinction between the high abuse liability profile of stimulants such as amphetamine and the lower abuse liability of other stimulants such as caffeine. In a series of three studies, the subjective and discriminative effects of several stimulant drugs are being investigated that are widely sold via mail order and designed to imitate amphetamine-like stimulants. These so called "look-alike" stimulants contain caffeine alone or combined with one or more sympathomimetic amines, such as ephedrine and phenylpropanolamine (PPA). Little is known about the behavioral pharmacology in humans of the combined effects of these drugs, including effects when individuals exceed the therapeutic dosage in an attempt to achieve amphetamine-like euphoria. These studies, which have recently been completed, are investigating the drugs singly and in combination. Another series of studies in progress is attempting to refine the drug self-administration paradigm to allow more systematic evaluation of the reinforcing effects of opioids. The initial study of opioid self-administration found that as drug dose was increased, responding for drug increased and concurrent responding for money decreased. The results of this study can be extended to evaluate the role of behavioral factors and medications in modulating the reinforcing effects of opioids in future research.